资源说明：In this thesis we present an operational computer vision system for real-time detection and tracking of human motion. The system captures monocular video of a scene and identifies those moving objects which are characteristically human. This serves as both a proof-ofconcept and a verification of other existing algorithms for human motion detection. An approach to statistical modeling of motion developed by Y. Song is coupled with a preprocessing stage of image segmentation and point feature tracking. This design allows a system that is robust with respect to occlusion, clutter, and extraneous motion. The results of experiments with the system indicate the ability to minimize both false detections and missed detections.

联系我们：verysource_com

CopyRight © 2008-2022 verySource.Com All Rights reserved.　京ICP备17048824号-1　京公网安备：11010502034788